Does Substack penalize fully free spaces?

Hypothesis-driven and data-informed results from a five-month, real-world experiment

Over the last months I treated Substack like a black box: apply one controlled perturbation (introduce a paid tier without changing the editorial rhythm), then watch how the system responds. The result? A clean signal for regime change in growth, plus unexpected patterns. What moved, what didn’t, and what the data can(not) tell about this platform?